At 03:14 PM 10/14/97 -0000, Mark Rauterkus wrote:
>
>FWIW, the which element that is a part of majorcool is, IMHO, on the mark
>for nearly all of the things I'm expecting to encounter.
For those that do not know the MajorCool internals, the default "wh*ch" as
Mark is referring is incredibly brain-dead. Here's the deal:
1. On the opening screen, the user is prompted for some value. Depending on
what "siteaddr" module is installed, this prompt could be for a login-id,
X.500 attribute, full name, or (as used by 99.99% of all MajorCool
installations) a fully-qualified RFC822 address.
2) The prompted value is then passed to a site-defined "siteaddr" module to
process this value (using whatever lookup methods are defined in the
module) and return a regular expression string that is intended to
approximate all possible permutations of that address with site-specific
transformations applied. For the default 99.99%, no regexp is used, but
rather a fixed string -- the same fully-qualified address -- is returned.
3) The "wh*ch" command as simulated by MajorCool uses either "=~ /$regexp/"
or "eq $fixed-string" to show a list of all matched addresses. In the case
of the fixed-string comparison, $mungedomain is not yet supported, although
I will probably add that eventually.
So, you can see that for the 99.99% majority of installations, the "wh*ch"
command in MajorCool is just an exact match. No partial substrings at all.
And he feels this is sufficient for most. Building on that, what if (for
Mj2) "wh*ch" would perform an exact match search on a properly constructed
RFC822 address, with substring matches bounced to the admin for approval?
--bill
Follow-Ups:
|
|
